Blockchain-Enabled NextGen Service Architecture for Mobile Internet Offload
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Blockchain-Enabled NextGen Service Architecture
for Mobile Internet Offload
Raman Singh Hitesh Tewari
School of Comp Sci & Stats School of Comp Sci & Stats
Trinity College Dublin Trinity College Dublin
Dublin, Ireland Dublin, Ireland
Thapar Institute of Engineering & Technology htewari@tcd.ie
Patiala, India
raman.singh@thapar.edu
arXiv:2104.08495v1 [cs.NI] 17 Apr 2021
I. A BSTRACT Today the world is using mobile cellular technologies such
as 3G, 4G, 4G-LTE and 5G with varying data transfer speeds.
The amalgamation of different generations of mobile cel-
We believe that an improved user experience can be gained
lular networks around the globe has resulted in diverse data
by offloading such cellular network users to local, higher
speed experiences for end users. At present there are no
speed networks. For example, if WiFi providers could allow
defined mechanisms in place for a subscriber of one mobile
mobile subscribers to use their fixed broadband infrastructure
network operator (MNO) to use the services of a WiFi
and in return get monetary reward for their services from a
provider. Cellular and Data Service providers also have no
MNO, then one can reduce the load on the cellular networks
standardized procedures to securely interact with each other,
and simultaneously increase the data speed offered to users.
and to allow their subscribers to use third party services on a
The proposed framework allows independent private WiFi
pay-as-you-go basis. This paper proposes a blockchain-based
operators to be paid for their services by offloading users onto
offloading framework that allows a subscriber of a mobile
their networks, and that in turn means less capital expenditure
operator to temporarily use another MNO or WiFi provider’s
investment by the MNOs. Some studies like the one released
higher speed network. Smart contracts allow diverse entities
by OpenSignal [2] suggest that in the future, mobile Internet
such as MNOs, Brokers and WiFi Providers to automatically
will transcend WiFi speeds, then offloading can also be
execute mutual agreements to enable the utilization of third
performed from WiFi to mobile network infrastructure. The
party infrastructure in a secure and controlled manner. To test
second reason for offloading can be the guarantee of services
the proposed framework, the offloading of a subscriber from
to subscribers by the MNO where they do not have a license to
3G/4G/4G-LTE/5G networks to a fixed broadband WiFi net-
operate, or have poor signal coverage issues. These subscribers
work was carried out and the results analyzed. The offloading
can be offloaded to partner WiFi providers and the subscriber
framework was implemented using the ns-3 network simulator,
will be able to enjoy enhanced data speeds. The third rationale
and the Ethereum blockchain smart contract features were used
for offloading is to ensure better service while roaming. For
for the settlement of invoices.
example, a subscriber who does not have roaming enabled on
their device, but wants to use the Internet for short periods of
II. I NTRODUCTION
time can be offloaded to one of these high-speed networks.
The global rollout of 5G networks is now gathering mo- To allow subscriber roaming between operators and the
mentum, as more and more countries are deploying this state- settlement of usage charges, MNOs at present have memo-
of-the-art broadband cellular network technology. However, at randa of understanding (MoUs) drawn up between them on
the same time many countries still have operational legacy a bilateral basis. However, MoUs are complex agreements
mobile networks such as 4G, 4G-LTE, or even 3G. Even in which take time to negotiate, and therefore it would not be
the places where 5G networks are available, coverage is not practical to negotiate MoUs with large numbers of small
always universal, and many pockets exist that still run older and medium sized WiFi providers on a per MNO basis. To
generation networks (e.g. 2G/3G/4G). A report published by enable the offloading process, MNOs can register with an
the GSM Association (GSMA) [1] suggests that at the end of intermediary/broker who can collectively negotiate MoUs with
2019, 4G coverage was about 50% of the total mobile Internet many WiFi providers on their behalf. Additionally, blockchain
availability by geographical area. Due to several advantages technologies can be used to allow the participating entities to
such as lower cost and fixed broadband/fiber infrastructure, trust each other by executing smart contracts on a blockchain.
WiFi still provides higher bandwidth speeds to its users, and This append-only distributed ledger technology (DLT), along
is popular among small and big organizations and also in retail with a consensus mechanism allows the implementation of
and residential settings. smart contracts in real terms, and also digitally facilitates,
Fig. 1. System architecture of the proposed framework
verifies and enforces the contract between two or more parties allows various other blockchain nodes to interact with each
[3]. Smart contracts can act as a bridging gap between stake- other, and to update the ledger periodically, including, adding
holders, and can provide subscribers with a new level of user or executing new smart contracts or transactions. The MNO
experience. authentication module helps to identify and authenticate a
subscriber from the MNOs subscriber database. Since there
III. B LOCKCHAIN - BASED S UBSCRIBER O FFLOADING
are many MNOs which are registered with the Broker, and the
F RAMEWORK
subscriber should be an active user of that particular MNO, this
Our proposed offloading framework enables subscribers module identifies a subscriber from an open smart contract,
to temporarily offload their data usage from low-bandwidth and authenticates its status in order to verify that the user a
to high-bandwidth channels, without changing their network valid subscriber and is authorized to offload.
operator. Fig. 1 represents the block diagram of the framework
and consists of three primary entities, namely a Broker, MNOs The smart contract module interacts with open smart con-
and WiFi Providers. The Broker as the name suggests is tracts to identify users, and sets the values of the parameters
the ingress of the whole process and is able to coordinate in the contracts based on the authentication status, such as
activities between all the entities, as every entity in the system success or failure. This module can access the smart contract’s
is registered with the Broker. We believe that a GSMA [4] like data based on the authorization allowed and helps in executing
entity aptly fits the role of the Broker in our proposed system, it. The billing module can access the transactions stored in
and all MNOs must register with it. The Broker maintains a the blockchain, and create a bill based on the executed smart
blockchain node along with a registration service. The MNO contract which involves a particular MNO. This module then
registration process includes setting up a blockchain node for matches the billing amount from the invoice received from the
storing smart contracts and transactional data. Organizations WiFi Provider and authorizes the payment.
that wish to allow their high-speed wireless infrastructure to be The third set of entities are the WiFi Providers which open
used by MNO subscribers must also register with the Broker, up their infrastructure in a controlled manner to the subscribers
along with setting up their corresponding blockchain node. The of MNOs. To expedite the offloading process, a WiFi Provider
Broker has oversight during the settlement phase, and also acts maintains a blockchain node on its premises. This entity also
as a mediator in the case of any disputes. operates on four modules, such as the blockchain interface,
The second set of entities in the system are the MNOs that authentication module, smart contract module and billing mod-
allow their subscribers to opt for offloading to a higher speed ule. The blockchain interface is responsible for maintaining
network. Reasons for offloading can include low-quality signal an up-to-date smart contract and transaction data, along with
coverage, high-speed requirements, roaming to non-serviced the full blockchain. The local authentication module ensures
areas, or even accessing services provided by particular Data the mobile number ownership of the subscriber by validating
Service providers. This entity includes various functionalities a one-time password (OTP). This local authentication of the
like a blockchain interface, authentication module, smart con- mobile number also avoids the spamming of mobile users
tract module, and a billing module. The blockchain interface or blockchain data. For example, if the local authentication
of a mobile number is not concluded successfully, spammers also generates a random one-time password (OTP MNO) and may create millions of offload requests using random mobile sets the value of OTPCheckMNO to OTP MNO in the smart numbers, which in turn would create a corresponding number contract. OTP MNO is also forwarded to the subscriber’s of open smart contracts, and force denial of service attacks mobile number for further processing. Once the subscriber against legitimate users. The smart contract module allows receives the OTP MNO value, it enters it on the landing page the WiFi Provider to create a new smart contract and set its of the WiFi Provider. The WiFi Provider in turn assigns the variables based on the subscriber authentication mechanism. OTP MNO value to the OTPCheckUser field of the smart Once the subscriber is allowed to offload and subsequently contract. Now, if the values of Auth WP and Auth MNO terminates its connection, the billing module records the data are both equal to 1, it deduces that both the WiFi Provider usage and writes this as a new transaction to the blockchain. and MNO have validated the subscriber’s identity. If the OT- Fig. 2 details the various phases of the offloading framework. PCheckMNO and OTPCheckUser are the same, it represents Phase 1 - One-time Registration: The registration process that the subscriber is authorized to offload, and the smart for a new entity such as a MNO or WiFi Provider wishing contract has been executed on the blockchain. The smart to join the system commences with the submission of their contract can also include other information like total time trusted third party (TTP) issued public key certificate to the allowed to offload, or any other conditions with the associated Broker. The Broker which maintains its own blockchain node offload which need to be honored by all the parties. stores the certificate as a transaction on the blockchain. All Phase 4 - WiFi Access Facilitation: In phase 4, the WiFi the communication amongst the entities in the system is Provider checks the status of the smart contract. If the contract carried out using public-key cryptography, and backed by is executed, the WiFi Provider allows access of its services transparent logs to ensure auditing at a later stage [5]. MNOs to the subscriber provided by its infrastructure. Primarily this and WiFi Providers join the blockchain network by creating service is high-speed Internet, but it can also be a wide their corresponding blockchain nodes. range of other services offered by WiFi providers. When Phase 2 - Offloading Initiation and Local Authenti- the subscriber terminates the connection, the WiFi Provider cation: In phase 2, an offloading request is initiated by a records the data consumed by the subscriber in its logs. user/subscriber. Once the subscriber is in the range of a Phase 5 - Blockchain Transactions and Billing: Phase WiFi Provider and wants to request an offload onto their 5 deals with the transactions and billing-related procedures. network, the subscriber connects with the wireless access point The WiFi Provider creates a new transaction in relation to (WAP) and accesses the landing page of the WiFi Provider. the data consumed by the subscriber and broadcasts it to the The landing page can be common for internal and external blockchain network. This transaction will be verified by all users. Providing their mobile number in both the username other blockchain nodes, and once verified, will be added to and password fields indicates that the user is external and a block by utilizing the proof-of-authority (PoA) consensus wishes to initiate an offload procedure. The user also selects mechanism by one of the authorized entities in the system. The its associated MNO from the drop-down menu provided on the PoA algorithm is able to provide faster transaction throughput landing page. The WiFi Provider generates the random one- using a “identity-as-a-stake” based consensus mechanism. It time password (OTP WP) and sends it to the mobile number significantly increases the speed of validating the transactions entered into the landing page by the user. The user then enters by generating blocks in a predictable sequence, and hence OTP WP on the next field of the landing page. Once the WiFi achieves a better transaction rate when compared with PoW Provider tests the validity of the mobile number, it generates a or PoS. The invoice can be generated by the WiFi Provider new smart contract and sets the value of Auth WP as 1. The after an agreed period of time such as a week or a month. WiFi Provider then encrypts the mobile number of the user The WiFi Provider will access all the transactions made by with the public key (PK MNO) of the mobile operator which it it from the blockchain data and prepare an invoice based on obtains from the blockchain, and assigns this encrypted value the mutually agreed per unit price. This invoice will also be to the field given in the smart contract. The WiFi Provider verified by the associated MNO and can also be ratified by the also includes the identity of the concerned MNO in the smart broker. Once all parties verify the invoice, the bill will be paid contract, so that all other entities know for whom the smart using a mutually agreed out-of-bounds payment mechanism. contract is intended. The Ethereum blockchain supports the scalability for large Phase 3 - MNO Level Authentication and Smart Con- scale offloading requests/transactions using a combination of tract Execution: Phase 3 describes the process undertaken sharding and side-chains. To test the performance of the by a MNO. The open smart contracts stored on blockchain Ethereum blockchain, researchers measured 4 million trans- are searched by the MNO. Once it finds the intended contract actions for 380 hours [6]. The experiment concluded that by matching the subscriber’s operator name, it will fetch the throughput decreases whereas latency increases linearly if we encrypted identity of the subscriber and decrypt it using its increase the block period which is fixed as per the difficulty private key. The decrypted data reveals the mobile number level of proof-of-work (PoW). As in our proposed framework, of the subscriber. The MNO tries to verify the identity of the we are using PoA strategy, this bottleneck should not affect subscriber against its subscriber database, and if successful sets the overall throughput and latency of the proposed system. the value of Auth MNO to 1 in the smart contract. The MNO To decrease the time required to complete the workload, the
Fig. 2. Sequence Diagram Illustrating the Various Phases
study suggests that powerful machines with high memory and implemented on the Docker [9] platform. Each node in the
CPUs should be used as blockchain node in PoA mode, for ns-3 network is connected to a Docker container using the
example, the computation time for workload can be reduced tap-bridge arrangement of ns-3 [10]. In the simulation environ-
by 25% if the memory is increased from 4GB to 24GB. If the ment, the subscriber is initially connected to the MNO node,
network size is considered, it is found that in 90%-100% of and when the smart contract is executed the connection is
cases, matches are found for smaller network sizes, whereas switched over to the WiFi Provider. The simulation was carried
the match ratio is merely 60%-75% for larger networks. out for 350 seconds on an Ubuntu Linux based computer
running a virtual machine with 8GB RAM, Intel i5 2.50 GHz
IV. C ASE S TUDIES AND R ESULT A NALYSIS processor, and 100 GB of allocated memory.
The proposed framework was implemented using the ns- The experimentation was conducted over five separate case
3 network simulator [7]. Various ns-3 nodes were created studies. The first case study takes the global average of Internet
to simulate the different entities e.g. Subscriber, MNO, speed and latency for fixed broadband and mobile Internet.
WiFi Provider, Broker and a Data Server. To implement The bandwidth given for the fixed broadband is assigned to
the blockchain functionality, the Ethereum [8] blockchain is the WiFi Provider link, whereas the bandwidth given for themobile Internet is assigned to the MNO link. In this case when compared to other case studies. As the data transfer
study, the subscriber is offloaded from mobile Internet to fixed speed decreases in the 4G and 3G case studies, we can
broadband as per the speed suggested by the global average. In see an increase in packet drop ratio for non-offloaded and
the subsequent case studies, the subscriber is offloaded from MNO flows. If these flows are offloaded to the WiFi network,
3G, 4G, 4G-LTE, and 5G mobile Internet to fixed broadband the packet delivery ratio rises to a better quality of service
i.e. the WiFi Provider. The various data transfer speeds and requirement.
latencies considered for all case studies are provided in Table
I.
TABLE I
C ASE S TUDIES AND A SSOCIATED PARAMETERS
Case Studies Network Generation Average Speed (Mbps) Latency (ms) Reference
Case Study 1: Fixed Broadband 92 21 [11]
Global Average Mobile Internet 46 36 [11]
Fixed Broadband 241 13 [11]
Case Study 2: 5G 71 20 [11]
Comparative 4G-LTE 50 50 [12]
Average 4G 10 100 [12]
3G 1.5 500 [12]
Fig. 3 shows the packet delivery percentages for all case
studies. The packet delivery is analyzed for all types of flows,
such as when the subscriber’s packets are not offloaded, for
offloaded packets, for packets transmitted through the WiFi
link, and other packets of different users which are being Fig. 4. Total Flow Duration Analysis For Offloading and No-offloading
transmitted through MNO network.
Fig. 4 exhibits the time taken to deliver a 500MB file, and 10
such total requests are made for each WiFi and MNO network.
One request of transferring a 500MB file is then offloaded
to a high-speed network. From this figure, it is evident that
the time taken to transfer files is significantly reduced in the
case of the offloaded flow. In the global average case study,
the non-offloaded flow takes 10.23 seconds to transfer one
file, whereas it is reduced to 0.91 seconds if the request is
offloaded. Similarly, the graph shows a drastic reduction of
delivery time in all other case studies. The longest time is
taken by the 3G network which is 339.11 seconds to deliver
the file, whereas it is reduced to only 0.49 seconds if this flow
is offloaded to the WiFi network.
Fig. 3. Packet Delivery Analysis for Various Types of Flows
For the global average case, the packet delivery percentage
is the same for all cases except for the MNO flows, however
this difference is insignificant. In this case, 99.96% of packets
are delivered for non-offloaded flows, offloaded flows, and
WiFi flows, whereas 99.94% of total packets are delivered for
the MNO flows. For all other case studies, it is evident from
Fig. 3 that 100% of offloaded flows are delivered, primarily
because of the enhanced data speed of the WiFi link. A slightly Fig. 5. Delay and Jitter Sum Analysis for Offloading and No-offloading
less number of packets are delivered in the global average
case study if compared to all other case studies, because Fig. 5 presents the analysis of delay sum and jitter sum for
the fixed broadband speed of the global average is lower all case studies. The delay sum is the addition of all delays foreach packet for the full duration of flow, whereas the jitter sum same offloading framework can be investigated on automated
is the addition of all jitter for every packet for any particular switching of network traffic from high-congested channels
flow. For the global average, the delay sum is calculated as to low-congested channels. This automated and agent-based
598.50 seconds. For other case studies such as 5G, 4G-LTE, framework could support load balancing and optimization of
4G, and 3G it is computed as 390.32, 806.39, 1749.86, and next-generation network infrastructure.
8466.05 respectively for the non-offloaded flows. If compared
R EFERENCES
to the delay sum obtained by the offloaded flows, we can see
a drastic reduction. The delay sums obtained for the offloaded [1] K. Bahia, and A. Delaporte, “The State of Mobile Internet Connectivity
2020”, https://www.gsma.com/r/wp-content/uploads/2020/09/GSMA-
flows are 45.34, 16.95, 16.75, 16.42, and 16.42 for global State-of-Mobile-Internet-Connectivity-Report-2020.pdf, Accessed on
average, 5G, 4G-LTE, 4G and 3G. 20-01-2021.
Jitter sum is also improved in case of offloading, whereas [2] I. Fogg, “The State of Wifi vs Mobile Network Experience
as 5G Arrives”, https://www.opensignal.com/sites/opensignal-
it is high if no offloading is performed. The jitter sum of 1.51 com/files/data/reports/global/data-2018-11/state of wifi vs mobile
seconds is reduced to 0.53 seconds for the global average case opensignal 201811.pdf, Accessed on 20-01-2021.
study if offloading is performed. It is reduced to 0.17 seconds [3] S. Wang, L. Ouyang, Y. Yuan, X. Ni, X. Han and F. Wang, “Blockchain-
enabled smart contracts: architecture, applications, and future trends.”
from 3.16 for 5G, 0.23 seconds from 1.41 seconds for 4G- IEEE Transactions on Systems, Man, and Cybernetics: Systems 49, no.
LTE, 0.28 seconds from 7.98 seconds for 4G, and impressive 11 (2019): 2266-2277.
0.28 seconds from 45.63 seconds for the 3G case study if [4] Global System for Mobile Communication Association (GSMA),
https://www.gsma.com, Accessed on 20-01-2021.
offloading is performed. The enhanced packet delivery ratio, [5] O. Gasser, B. Hof, M. Helm, M. Korczynski, R. Holz, and G. Carle, “In
along with improved delivery time, reduced delay sum and log we trust: Revealing poor security practices with certificate trans-
reduced jitter sum, enhance the overall quality of experience parency logs and internet measurements.” In International Conference
on Passive and Active Network Measurement, pp. 173-185. Springer,
provided by the MNOs to their subscribers. Cham, 2018.
[6] M. Schäffer, M. Angelo, and G. Salzer, “Performance and scalability of
V. C ONCLUSIONS private Ethereum blockchains”. In International Conference on Business
Process Management, pp. 103-118. Springer, Cham, 2019.
In this paper, an offloading framework is presented with [7] ns-3 a discrete-event network simulator, https://www.nsnam.org, Ac-
its several advantages and infrastructural benefits. It can fa- cessed on 20-01-2021.
cilitate MNOs to allow their subscribers to benefit from third [8] Ethereum, https://ethereum.org/en, Accessed on 20-01-2021.
[9] Docker Container, https://www.docker.com, Accessed on 20-01-2021.
party operator high-speed infrastructure for a particular time [10] R. Singh, “StepByStep: Establishing virtual net-
period. To the best of our knowledge, there are no such work between Docker container and NS-3 nodes”,
mechanisms that exist as of now which can support such https://sites.google.com/thapar.edu/ramansinghtechpages/step-wise-
establishing-connection, Accessed on 20-01-2021.
inter-organizational arrangements securely and efficiently. Our [11] Speedtest Global Index, https://www.speedtest.net/global-index, Ac-
proposed framework deploys smart contracts for authentica- cessed on 20-01-2021.
tion, thereby allowing the subscriber to offload, and utilizes [12] L. Rizzatti, R. Squiers and M. Castren,“Design and verify 5G systems,
part 1”, https://www.edn.com/design-and-verify-5g-systems-part-1/, Ac-
blockchain transactions to record and generate invoices. The cessed on 20-01-2021.
MNO and private WiFi Provider both authenticate the sub-
scriber and its mobile number to rule out any spamming of the
system. All the transactions are verified by each participating
entity, and new blocks are added using a PoA consensus
mechanism to minimize the mining effort.
The proposed framework was simulated using ns-3, and
the Ethereum blockchain was integrated into the simulation
environment using Docker containers. A total of five case
studies i.e. global average speed, 5G, 4G-LTE, 4G, 3G to WiFi
offloading were tested. The final analysis shows that offloading
results in improved packet delivery ratios, reduced total flow
duration, total delay, and total jitter. These parameters suggest
that offloading can help in enhancing the end users quality of
service experience. A present, Internet speeds vary geograph-
ically as well as amongst operators. A user has no choice but
to switch operator if they need higher speeds, or services that
cannot be provided by their operator. Our proposed offloading
framework can be a great leap forward for subscribers who
can enjoy higher bandwidth speeds on a on-demand basis.
In the future, challenges such as the scalability of
blockchain transactions, simultaneous subscriber load, etc. can
be analyzed. Time lag analysis of the high load of subscriber
offloading requests can also be carried out. In addition, theYou can also read
